Potential Adverse Effects of Proton Pump lnhibitors in the Elderly

Ami Kapadia, MD, Daisy Wynn, MD, and Brooke Salzman, MD

Introduction dence of side effects is similar to placebo, which is less than Since the introduction of omeprazole in 1989, proton 5%.’0 The most common side effects reported are headache, pump inhibitors (PPIs) have become one of the most com- diarrhea, abdominal pain, and nausea. The etiology of these monly prescribed classes of medications in the world. In side effects, particularly diarrhea, may be related to alter- 2007, PPI sales in the United States ~vere in excess of$1 ! bil- ations in gut flora caused by acid suppression50 The only lion5 Esomeprazole and lansoprazole both ranked among contraindication to the use of PPIs is a known history of the top five drugs sold in the United States in 20075 hypersensitivity to this class of medications. PPIs do not Overall, with their high safety profile and demonstrated require dosage adjustment in renal or hepatic dysfunction. efficacy, PPIs represent a major advance in the treatment of acid-related disorders ranging from peptic ulcer disease to Indications and Approved Uses erosive esophagitis. However, it has been shown that PPIs are PPls are accepted for the treatment and remission mainte- often misused and overused, which may have significant nance of acute gastric and duodenal ulcers, symptomatic gas- -7 implications? With the widespread and frequent long-term troesophageal reflux disease (GERD), erosive esopha~gitis, use of PPls, several adverse effects have come to light that pathological hypersecretory conditions such as Zollinger- may call for more selective prescribing practices, particularly Ellison syndrome, nonsteroidal anti-inflammatory drug- in older adults who may be more vulnerable and likely to induced gastric ulcers, and eradication of Helicoba/terpy/ori suffer the consequences of such adverse effects. With an esti- (Table I). mated 8% of males and 15% of females age 65 years and In 2008, the American College of Foundation older experiencing reflux and potentially using acid-suppres- Task Force on Clinical Expert Consensus Documents pub- sive therapy,8 understanding the risks for potential adverse lished recommendations regarding the utilizaqon of PPIs to effects associated with PPls is critical in this population. decrease the risk of upper-gastrointestinal (GI)’ events associ- In this article, we review the current data on selected neg- ated *vith antiplatelet therapy for primary or secondary treat- ative outcomes that may result from PPI use. Specifically, ment of cardiovascular disease54 The recommendations sug- increasing evidence demonstrates that PPI therapy may be gest that PPIs should be used in the following clinical situa- associated with the development of Clostridium di~cile Conllkqo tions in order to reduce risks of GI bleeding: when patients , hip fractures, community-acquired pneumonia, ~vhich la are taking aspirin or dopidogrel and have a history of an ment of vitamin B12 deficiency, and possibly immunoglobulin E- ~ . ulcer (either bleeding or nonbleeding); or when p~tlents are "stress" mediated allergic reactions. The implications of such adverse taking dua! antiplatelet therapy or concomitant anficoagula- outcomes, along with the evidence of the inappropriate use teroids tion therapy24 of PPIs, underscore the need for more judicious use of this class of medications. inpatien Overuse/Misuse Drug Pathophysiology While it is clear that PPI therapy can be beneficial ~vhen used appropriately in the treatment of various GI disorders, and/or PPIs achieve acid suppression by binding to a specific evidence suggests that PPls are often overused and misused?r not clini { enzyme on the parietal cell, known as the proton pump, In many cases, PPIs are initiated or continued for prolonged resulting in inhibition of gastric acid secretion by the proton tematic 9 periods of time, without sufficient evaluation of the need for pump. PPIs are generally well tolerated. The overall inci- therapy. In fact, Manoharan et al,5 showed that in more than polyphal sequenc{ half of elderly inpatients who were prescribed acid-suppres- affect th. Dr. Kapadia is a Family Medicine , Grand Junction sion therapy upon hospital discharge, there was no clearly COll~lnol VA Medical Center, Grand Junction, CO; Dr. Wynn is a Fami- documented indication for such therapy. Another study by ly Medicine Physician, Community Health Center of Meriden, 4 nifedipii Meriden, CT; and Dr. Salzman is Assistant Professor, Nardino and colleagues showed that 65% of medical inpa- affect th tients who xvere prescribed PPIs did not have an approved Department of Family and Community Medicine, Division of tion, PP Geriatric Medicine, Thomas Jefferson University Hospital, indication for use. Yet another study conducted at a major system a Philadelphia, PA. Australian teaching hospital showed that 63% of medical also usi~: inpatients were receiving PPIs for unapproved indications? greatest 24 Clinical July/August 2010 FDA-Approved Indications for PPI Use and Duration of Therapya

Indications Duration of Therapy Acute gastric and duodenal ulcers 4-8 wk Maintenance of healed duodenal ulcer Unknown Symptomatic GERDb 4 wk (omeprazole) or 8 wk (lansoprazole) Erosive esophagitis 4-8 wk Maintenance of erosive esophagitis Unknown Zollinger-Ellison syndrome Several yrs NSAID-induced gastric ulcer prophylaxis Duration of NSAID NSAID-induced gastric ulcers 4-8 wk Eradication of Helicobacterpylori 7-14 days with tr pie antibiotics (meta-analyses and controlled studies have indicated littie clinical utility beyond 7 days)

FDA = Food and Drug Administration," PPI = proton pump inhibitor; GERD = gastroesophageal reflux disease; NSAID = nonsteroidal and-inflam- matory drug, a Contains information from references 10 and 12. Atypical symptoms of GERD, such as early satiety, odynophagia, and weight loss, should be investigated with an upper endoscopy prior to eropiric initiation of an acid-suppressive agent,

Effects of Concomitant Use of PPIs with Commonly Prescribed Medications in the Elderlya

Drug Affect of PPI on Metabol sm Chmcal Relevance Digoxin ’ Increases absorption I Increase in digoxin levels and digoxin-associated toxicity , , Nifedipine Increases absorption Hypotension, shock, coma Warfarin Decreases metabolism Elevation in INR, b eeding

PPI = proton pump inhibitor; INR = international normalized ratio. Contains information from references 10, 17-20.

Common examples of unapproved utilizations of PPIs, studies have shown that the ~nmadon of omeprazole can which lack evidence to support their practice, involve treat- increase the absorption of digoxin, resulting in toxic digoxin ment of low-risk inpatients as a means of prophylaxis against levels,’8’~9 and can decrease the metabolism of warfarin,’7 "stress" ulcers and concurrent administration with corticos- resulting in an increased international normalized ratio reroids to prevent peptic ulcer disease in noncritically ill (INR) and bleeding,z° inpatiems)-7 In January 2009, the Food and Drug Administration released an early communication about the potential interac- Drug Interactions and Polypharmacy tions between omeprazole and clopidogrel.~’ The concern I)o/)~/)/Jarr,~ac;y, defined as the use of multiple medications that omeprazole may lower the efficacy of clopidogrel and/or the administration of multiple medications that are remains controversial and unsubstantiated given that avail- not clinically indicated,~6 is a common and potentially prob- able studies show conflicting results.22-25 However, a recent lematic occurrence in the elderly. One consequence of large, retrospective cohort study by Ho et a126 found that polypharmacy is drug interaction. This is an important con- patients taking a PPI and clopidogrel post-acute coronary sequence to consider when initiating PPIs since they can syndrome have a modest increased risk of death or rehospi- affect the absorption and metabolism of several medications talization as compared to those patients taking clopidogrel commonly prescribed in the elderly, such as digoxin, alone (odds ratio [OR], 1.25; 95% confidence interval [CI], nifedipine, and warfarin,l° By altering gastric pH, PPIs can 1.11-1.4). The results of this study suggest that the risks and affect the bioavailability of different medications. ~7 In addi- benefits of PPI therapy in patients taking clopidogrel should tion, PPls are metabolized by the hepatic cytochrome P450 be carefully considered. Further research studies released in system and thereby can affect the metabolism of medications 2010 continue to show conflicting results.27’’-8 Overall, there also using this system (Table II). Omeprazole possesses the are strong pharmacodynamic data suggesting that an inter- greatest potential for altering the P450 system. For instance, action exists between clopidogrel and PPIs; however, the

Clinical Geriatrics July/August 2010 25 Although CDAD was more strongly associated with antibiot- c Overall i use (OR, 13.1; 95% CI, 6.6-26.1), the results suggest that ranging f There is enough evidence the risk of CDAD in hospitalized patients receiving antibi- ment of ( otics may be compounded by exposure to PPI therapy,e° to warrant concern about In 2004, Dial and colleagues~1 conducted a cohort study PPls a involving 1187 inpatients who received antibiotics. Within In addi the possible role of PPIs in this cohort, patients who had also received a PPI or H2 a role in blocker were compared to patients ~vho had not received any on acid-sl patients at risk for CDAD. form of acid-suppressive therapy. Overall, 81 (6.8%) of the possibly 1187 inpatients developed CDAD. CDAD was significantly severe all, associated with use of PPIs (adjusted OR, 2.1; 95% CI, 1.2- nificant clinical significance of this interaction is still the subject of 3.5).31’ The investigators also included a case-control study continued debate and ongoing research. the potet during the same time period in order to control for other (IgE)-m( Overall recommendations would suggest caution in pre- confounding factors that could not be addressed in the scribing PPI therapy to patients who also take clopidogrel, cohort study. In this case-control study, cases were identified increased: but acid suppression should not be ~vithheld if there is a clear as patients with a history of diarrhea and a positive C. acidic co~ indication for its use. Clinicians should also be mindful that die toxin assay. Controls were chosen from a list of patients This in the elderly, PPI use with digoxin, nifedipine, and warfarin who had been prescribed antibiotics ,vhile in the hospital can potentiate digoxin toxicity, hypotension, and increased undTsto during this same study period. Control subjects were gasmc p INR, respectively (Table II). matched to cases by inpatient ward, age, class of antibiotics, pH of and, ,vhen possible, number of antibiotics. Overall, it was Therefo, PPIs and Clostridium difficile-Associated found that CDAD was significantly associated with the use Diarrhea digestim of PPIs (adjusted OR, 2.7; 95% CI, 1.4-5.2)2’ ed in th~ Ulostridium di55%ileassociated diarrhea (CDAD) is the A more recent case-control study by Aseeri et aD also been most commonly diagnosed hospital-acquired diarrhea in showed that PPI use was a risk factor for developing CDAD is a prer{ developed countries. It occurs with an incidence of 0.1-2% in hospitalized patients. The investigators pair-matched 94 in hospitalized patients and has been reported to be increas- the food patients who developed CDAD with control patients, con- An ok ing in frequency and severity.29,3° Clinical symptoms can trolling for the following common predisposing risk factors: range from mild diarrhea to serious life-threatening cases of patients admission date; antibiotic exposure; gender; ~ge; patient blockers pseudomembranous colitis and colonic perforation. location (medical vs surgical unit); and room type. The antibod! Established risk factors for CDAD include antibiotic thera- investigators reported that patients with CDAD were signif- 3 montl py, advanced age, significant comorbidity, chemotherapy or icantly more likely than controls to receive acid-suppressive that 25c, other immunosuppressants, GI procedures and surgery, stay therapy during hospitalization: 76.6% versus 42.6%, respec- forap in an Intensive Care Unit, residence in a nursing home, and tively (P = 0.030). Multivariate analysis showed that CDAD 2~,3° toward prolonged hospitalization. More recently, it has been pro- was associated with the use of PPIs with an OR, of 3}6 (95% treated 29 posed that acid-suppressive therapy is also a risk factor for CI, 1.7-8.3; P < 0.001). order tc CDAD. Gastric acid has a protective role in the defense Finally, a recent cohort study conducted by Howell and ing the against ingested bacteria, and thus, decreasing acid would 36 colleagues looked at the intensity of acid-suppression ther- change lead to diminished defense mechanisms against foreign apy in relation to the risk of nosocomial CDAD. This study Of the invaders.29’3° Decreased gastric acidity is a known risk factor involved analysis of data collected on over 100,000 dis- formed for other causes of infectious diarrhea, including salmonella charged patients from a tertiary medical center over a five- ing foo~ and cholera.29’3° Having a higher gastric pH, such as the pH year period. The risk of nosocomial CDAD was examined achieved with PPI use, may facilitate colonization by C. d~ compar in four different groups, including those receiving no acid risk of ficile with subsequent toxin production and toxin-mediated suppression, histamine2-receptor antagonist (H2RA) thera- respons intestinal injury and inflammation.29 py, daily PPI therapy, and PPI therapy more frequently than Multiple studies have examined the possible role of PPI use lO.5 (5 daily. They found that as the level of acid suppression after di in the development of CDAD. Although conflicting data increased, the relative risk of nosocomial CDAD increased, reactivi exist involving this potential relationship, there is enough evi- from 1 (no acid suppression) to 1.53 (95% CI, 1.12-2.10; dence to warrant concern about the possible role of PPIs in H2RA), to 1.74 (95% CI, 1.39-2.18; daily PPI), and to results,patient: patients at risk for CDAD.2935 In a prospective case-control 2.36 (95% CI, 1.79-3.11; more frequent than daily PPI)26 study of 155 consecutive inpatients with CDAD, it was (includ Their results suggest that increasing levels of pharmacologic quent f shown that CDAD ,vas independently associated with acid- acid suppression are associated with increased risks of noso- ~vith e suppression therapy (OR, 1.90; 95% CI, 1.10-3.29).30 comial CDAD26 under~

26 Clinical Geriatrics July/August 2010 mtibiot- Overall, the data suggest an appreciable association (OR percentage of positive skin reactions was higher in exposed ~,est that ranging from 1.90 to 3.6) between PPI use with develop- patients versus controls (7.02% of patients vs 1.00% of antibi- ment of CDAD. controls had positive skin reactions). Overall, the relationship between PPI therapy and the rt study PPls and Food development of IgE food allergens remains unclear but may Within In addition to their relation to CDAD, PPIs may also play be significant. More studies in this area are needed to clarify or H2 a role in promoting increased allergic reactivity in patients this potential association, and whether there are meaningful red any on acid-suppressive therapy27 Rising food allergen formation clinical implications. of the possibly associated with PPIs may result in escalating rates of ficantly severe allergic reactions such as anaphylaxis and lead to sig- PPls and Community-Acquired ]I, 1.2- nificant restrictions in diet. Untersmayr et aP7 investigated Pneumonia i study the potential role of PPIs in promoting immunoglobulin E H2RAs and PPIs create a hypochlorhydric to achlorhy- : other (IgE)-mediated food allergies. Based on their animal experi- dric environment in the gut, thereby facilitating survival of in the ments, it was hypothesized that food proteins may cause certain ingested pathogens that would otherwise be killed :ntified increased aliergenicity if they survive through the normally by unaltered pH gastric juices. Regurgitation of ingested acidic contents of the stomach. bacteria into the oropharynx can precipitate respiratory ,atients This ~echanism of increased IgE reactivity can be infections. ospital understood by realizing that acid-suppressive therapy raises In 2004, Laheij and colleague9 published a prospective were gastric pH to around 5.0. It is known that the optimum cohort study with a nested case-control analysis of a 3iotics, pH of pepsins in the stomach is between 1.8 and 3.2. national health database in the Netherlands to examine the it was Therefore, acid suppression impairs the peptic enzymatic association between acid-suppression use and community- ixe use digestion of proteins2~ When food proteins are not degrad- acquired pneumonia (CAP). Of the 364,683 adults who ed in the stomach, they retain their protein structure. It has were studied over a period of approximately 2.7 years, :v a~so been speculated that this preserved food protein structure 19,459 were considered exposed, defined as receiving a first [DAD is a prerequisite for the development of IgE antibodies to prescription for an acid-suppressive agent. It was reported ~ed 94 the food.’~ that 10,177 received H2RAs and 12,337 received PPIs, , COI1- An observational cohort study ~vas performed with 152 while 3055 individuals used both. In the exposed group, lctors: patients who ~vere starting treatment with H2 receptor 477 individuals developed pneumonia, either while taking ,atient blockers or PPIs. Blood samples to test for IgE-specific acid-suppressive therapy (185) or after discontinuation of . The antigodies to 19 specific foods were drawn before and after the acid-suppressive agent (292). There were 5366 persons ;ignif- 3 months of acid-suppressive therapy. Investigators found with pneumonia in the nonexposed group. The relative risk essive that 25% of patients who were on acid-suppressive therapy for patients taking an acid suppressant developing CAP was espec- for a period of 3 months formed new IgE antibodies 1.6 as compared to those patients not taking an acid sup- DAD toward food antigens. Fifty cbntrol patients who were not pressant. Pneumonia was confirmed by chest x-ray, sputum (95% treated with acid-suppressive therapy were recruited in culture, or clinical symptoms. To reduce confounding by order to exclude a nonspecific boost in IgE reactivity dur- indication, a nested case-control analysis was conducted i and ing the pollen season. The control subjects showed no among the 477 adults who developed pneumonia during or ther- change in their IgE patterns during the course of the study. after discontinuation of the acid-suppressive medication. ~mdy Of the 25% of patients on acid-suppressive therapy who Four hundred seventy-five of those patients were matched dis- formed new IgE antibodies, 10% had a boost of pre, exist- to 4690 controls for sex, age, and index date of pneumonia. ing food-specific IgE, and 15% had new IgE formation. As Acid-suppressive use was defined as current if the prescrip- ~ined compared to 50 age-matched healthy controls, the relative tion included the index date of pneumonia, and was acid risk of developing the increase of an IgE food allergen ...... if the end of the last prescription was before hera- response after 3 months of acid-suppressive therapy was the index date. Past was further broken down into recent than 10.5 (95% CI, 1.44-76.48; 19 = 0.0203).3~ Five months past if the end of the last prescription was less than 30 days, after discontinuation of the H2 blocker or PPI, serum IgE past if between 30 and 180 days ago, and distant past if tsed, reactivity decreased but was still present in 6% of more than 180 days ago. Laheij et a139 pointed out that as LIO; patients27 To determine the in vivo relevance of these compared to the controls, the patients with pneumonia d to ~’ results, skin prick testing was done. After 8 months more often had comorbid conditions such as mel- ’I) 2 (including 3 mo of acid-suppressive therapy and subse- litus, heart failure, and pulmonary disease, as well as antibi- ogic quent 5 mo after discontinuation of therapy), 33 patients otic usage. After adjustment of confounders such as med- OSO- with elevated serum, allergen-specific IgE antibodies ical comorbidities and use of antibiotics, it was found that underwent skin prick testing with the 19 food extracts. The current use ofa PPI or a H2RA was associated with a small

Clinical Geriatrics July/August 2010 27 increase in risk of CAP (adjusted OR, 1.27; 95% CI, 1.06- acid-suppressive medication, 83% received a PPI and 23% ratio [HR 1.54). Within this group, the risk of CAP appeared most received a H2RA. The adjusted OR of t-LAP in those receiv- studies,45-4 pronounced among patients who began an acid suppres- ing a PPI after adjustment of age and comorbidities was 1.3 be>,veen 1 sant within the last 30 days (adjusted OR, 2.12; 95% CI, (95% CI, 1.1-1.4). The adjusted OR of HAP in those receiv- ture risk; 1.29-3.48). Researchers also found a positive association ing a H2RA was 1.2 (95% CI, 0.98-1.14). The authors women between increased defined daily dose (DDD) of a PPI and found that the association between acid-suppressive agents did find a CAP: more than one DDD yielded an adjusted OR of 2.28 and HAP was only statistically significant for PPI use. CI, 1.15- (95% CI, 1.26-4.10). Overall, studies suggest that there is a modestly increased CI, 1.05- 4° In 2007, Gulmez and colleagues published a popula- risk of developing CAP related to current or recent use of 1.15-1.3~ tion-based, case-control study of patients in Denmark. PPIs. However, some feel that the association of PPIs with Based They studied 7642 patients with a discharge diagnosis of CAP may result from the confounding presence of the gesting tk CAP. CAP ~vas confirmed by x-ray, culture, and polymerase underlying condition, namely GERD, and not PPI use fi’acture i: chain reaction test. Current use of a PPI was defined as use itself.~3’43’44 Perhaps those patients with more severe reflux are risks relat in the past 90 days prior to CAP diagnosis, and_past use was at higher risk for aspiration and subsequent CAP regardless studied fi defined as greater than 90 days. Patients with CAP who of their use of an acid-suppressive agent.44 Further studies are may resul ~vere currently taking a PPI had a moderately increased risk needed to elucidate the nature of the relationship between ty contro of CAP (OR, 1.5; 95% CI, 1.3-1.7). The risk was most pro- the occurrence of CAP and PPI therapy. point out found in users who started a PPI within 0-7 days (OR, 5.0; ly than tt 95% CI, 2.1-11.7). Unlike Laheij et al,39 Gulmez et al did PPIs and Hip Fractures seizure a: not see an increased risk of CAP among H2RA users or in Recent studies have demonstrated an association between sal3tS, ma those taking increased daily doses of PPls. They did not find long-term PPI use and hip fractures in the elderly.45-~8 It has for falls, an association with recent past use (last prescription 90-180 been proposed that an acidic environment facilitates calcium ship betv days before CAP index date) or distant past use (> 180 days absorption in the gut% therefore, PPls can decrease calcium eratlon, r before CAP index date) of acid-suppressive agents. absorption. PPIs can also interfere with bone metabolism by ly impac~ { In 2008, Sarkar and colleagues4~ published a large (case weakly inhibiting proton transport within osteoclasts, and patients, n = 80,066; control participants, n = 799,881), possibly suppressing osteoclast ~tctivity.49 PPIs s nested case-control study of a United Kingdom database In 2006, Yang et al4~ published a large, nested case-control Anoth to examine the association between PPI use and CAP. study of patients from a United Kingdom database focusing col~alami Overall, they found that patients who began a PPI within on patients age 50 years and older who took a PPI for at least proposec tile past 30 days had an increased risk for CAP. This risk 1 year. They found a positive association between increashlg sorption increased dramatically in patients who began a PPI less duration and daily dosages of PPI and hip fracture. One year the relea: than 2 weeks prior to CAP diagnosis. Specifically, initia- of PPI use had an adjusted OR of 1.44 (95% CI, 1.26-1.54), potential tion of PPI 2, 7, and !4 days prior to diagnosis yielded an and 4 years of PPI use had an adjusted OR of 1.59 (95% CI, grmvth t OR of 6.53 (95% CI, 3.95-10.80), of 3.79 (95% CI, 1.39-1.80). PPI use of greater than 1 daily dose had an the smal 2.66-5.42), and of 3.21 (95% CI, 2.46-4.18), respectively. adjusted OR of 2.65 (95% CI, 1.80-3.90). ~ absorbec Patients with longer-term, current use of a PPI did not The same year, Vestergaard and colleagues~6 published a absorpti, have an increased risk for CAP. large case-control study performed in Denmark and found could re PPIs take approximately seven days to yield maximum an increased risk of any fracture in patients using a PPI with- cell prol 4’ ¯ acid-suppressive effect. It is hypothesized that aspiration of in 1 year. The OR of hip fracture with 1 year of PPI use was Occur Ill pathogens into the hypochlorhydric to achlorhydric envi- 1.60 (95% CI, 1.25-2.04). However, they were unable to acid sup ronment in the gut can predispose a patient to CAP. As confirm Yang et al’s finding of increased hip fracture risk with !or ~ Sarkar et aP noted, intuitively, one would predict that those with increasing daily doses of PPI. levels~,~ patients who have been taking PPIs chronically would be A more recent study published in 2008 by Targownick et patients the most at risk of developing CAP. However, these three a147 analyzing a Canadian database found that PPI use of 5 or opposed { 394~ studies failed to support this hypothesis. In contrast, the more years was modestly associated with increased risk of hip are still ~ results of these studies suggest that an increased risk of CAP fracture (adjusted OR, 1.62; 95% CI, 1.02-2.58). populati is associated with initiation and recent use of an acid-sup- In 2010, Gray and colleagues5° published results fi’om a cobalai~ pressive agent. prospective analysis of postmenopausal women enrolled in nificant Herzig and colleagues42 recently published the first the Women’s Health Initiative Observational Study and or neurt prospective study examining the relationship between acid- Clinical Trials taking a PPI or a H2RA for less than 1 year, with reic suppressive agents and hospital-acquired pneumonia 1-3 years, or greater than 3 years. The study did not show an With (HAP) in nonventilated patients. There were 63,878 increased risk in hip fractures in the 3396 women who were age of 6 admissions eligible for the study. Of the 52% prescribed an taking a PPI for any of the above lengths of time (hazard BI2 def

28 Clinical Geriatrics July/August 20!0 23% ratio [HR], 1.00; 95% CI, 0.71-1.40). Unlike the previous 4547 5° :ceiv- studies, Gray et al did not find a positive correlation It may be prudent to monitor ~s 1.3 between longer duration of PPI use and increased hip frac- ture risk; however, the authors noted that they had fewer ceiv- 812 levels in patients who women (392) taking a PPI for more than 3 years. The study gents did find an increased risk of spine fractures (HR, 1.25; 95% are chronically taking acid- CI, 1.15-13.6), forearm or wrist fractures (HR, 1.26; 95% ~ased CI, 1.05-1.51), and total fractures (HR, 1.25; 95% CI, suppressive therapy or in those se of 1.15-1.36) in women using a PPI. with Based on these aforementioned studies, the evidence sug- already at higher risk for B12 ~ the gesting the association of PPI use with increased risk of hip deficiency, such as the elderly use fractur~ is modest at best and remains unclear. The specific K are risks related to the duration and dose of PPls also need to be or those with dementia. dless studied further. The correlation of PPIs with hip fractures 2s are may result fi’om confounding factors that were not adequate- ween ly controlled. For instance, critiques of Yang et al’s study point out tl{at the patients with hip fractures were more like- adults is thought to be related to malabsorption of dietary ly than their controls to have used medications such as anti- B12. This malabsorption may be from atrophic gastritis, age- seizure agents, antipsychotics, anxiolytics, and antidepres- related hypochlorhydria, or other conditions such as H. pylori infection2 It can take months to years for a B12 defi- sants, many of which have also been linked with higher risks weell s~ ciency to manifest secondary to malabsorption because liver has for falls, osteoporosis, and hip fracture, While the relation- 9 ship between hip fractures and PPIs deserves serious consid- stores of B12 can compensate for decreased dietary intake. cium Valuck et aP examined the relation of acid suppression and dtlph eration, more definitive studies are still needed to significant- B12 deficiency by conducting a case-control study with ly impact prescribing practice. m by patients in a geriatric primary care setting. They enrolled and PPls and Vitamin B~2 Deficiency patients age 65 years and older who had documentation that Another proposed adverse effect of long-term PPI use is a serum BI2 laboratory assessment had already been complet- ntrol ed. They compared 53 vitamin B~2-deficient cases with 212 ising cobalamin malabsorption. Several mechanisms have been proposed by which PPI use may lead to cobalamin malab- controls for past or current use of acid-suppression therapy least based on subjects’ medical records. The following variables asing sorption. First, a more basic gastric environment ,nay slow r were controlled for: age; gender; multivitamin use; and H. year the release of cobalamin from dietary food sources.49Anothe potential mechanism involves the risk of bacterial over- i~ylori . Overall, researchers found that current use .54), of acid-suppression therapy for at least 12 months xvas asso- ~, CI, growth that can result from PPI use. The excess bacteria in ~he small intestine may consume cobalamin before it can be ciated with a significantly increased risk of vitamin B12 defi- ciency (OR, 4.45; 95% CI, 1.47-13.34).9 However, they did absorbed.49,5’- Yet another mechanism of decreased cobalamin not find an association between short-term current use or ted a absorption involves the theoretical possibility that PPIs could reduce intrinsic fiactor secretion by inhibiting parietal past use of acid suppression and Bp_ deficiency. )und ¯ Another study looked at the potential association between .vith- cell proton pumps; however, .this has not been found to occur in actual practiced Overall, studies have shown that chronic acid suppression and the initiation of B12 supple- 2 ’vVaS mentation. Force and colleagues5~ conducted a retrospective acid suppression may decrease vitamin B12 absorptio ie to n,,and case-control study, in which 125 cases were identified as risk with long-term use, may result in reductions in serum B12 levels.49,54 However, questions still remain regarding whether those patients who initiated vitamin B12 supplementation during the defined study period. The 125 cases were ck et patients may develop clinical vitamin B12 deficiency as opposed to simply exhibiting reductions in serum levels that matched to 500 controls. The cases were then compared to ?5 O17 49’54 the controls with regard to exposure to chronic acid-suppres- i’hip are still within the normal range. a particularly vulnerable population would be older adults, who are already at risk for sion therapy (defined as treatment with H2 blocker or PPI use for > 10 mo of the 12 mo that preceded initiation ofB 812 n )Vi~ a cobalamin deficiency. Vitamin deficiency can have sig- nificant consequences such as the development of dementia supplementation) .55 Overall, 23 cases (18.4%) were found to d in have used chronic acid-suppression therapy as compared to alld or neuropathy. Such conditions may not fully reverse, even with repletion.‘) 55 (11.0%) of the control group. Patients who initiated 5,ear, supplementation were more likely than controls to have \V an With regard to the geriatric population, adults over the age of 65 years are estimated to have a 5-15% prevalence of received chronic acid-suppression therapy (P = 0.025; OR, were 1.82; 95% CI, 1.08-3.09).55 ~zard B12 deficiency.‘) The etiology for BI2 deficiency in older Clinical Geriatrics July/August 2010 29 Conclusion and Level of Evidence of Potential Adverse Effects of PPI Use 9. Valuck ! ton pun Epidem~ Potential Adverse Conclusion Level of Evidence 10. Vanded" Effects of PPI Use , 2002;6~ CDAD 11. Pham C Significant associat on between PPI use and deve opment of CDAD (OR 1.90-3.6) Evidence level B, tient int~ case-control, cohort 12. Shi S, }~ (OR, 1.90; 95% CI 1.10-3.29~) ~’) studies cokineti (adjusted OR, 2.1; 95% C, 1.2-3.5 2008. (adjusted OR, 2.7; 95% CI 1.4-5 2~’) " 13. Detsky ’ (OR 3.6 95% CI 1 7-83~) munity-; 14. Bhatt E Increasing levels of acid suppression associated with increased risk of nosocomial Founda CDAD 2008 e antiplat~ OR 1 (no acid suppression) to 1.53 (95% CI, 1.12-2.10; H2RA), to 1.74 (95% CI, Cardiol~ 1.39-2.18; daily PPI), and to 2.36 (95% CI, 1.79-3.11; more frequent PPI)~ Circulat 15. Manoh~ CAP Current PPI use is modestly associated with an ncreased risk of CAP (OR, 1.27; Evidence level B, Ageing ’ 95% CI, 1.06-1.54~) (OR, 1.5; 95% C!, 1.3-1.74°) 16. Haijar E case-control studies Phartn~ Current PPI use initiated within the past 0-7 days increases risk of CAP 17. Sutfin’f (OR, 5.0; 95% Cl, 2.1-11.7’°); 2 days (OR, 6.53; 95% Cl, 3.95-10.80" ) 7 days wadarir (OR, 3.79; 95% Cl, 2.66-5.42~’) 18. Kiley C J2007; Food 19. Oosted PPI use may be associated with development of new!elevated IgE ant bed es to Evidence level C, foods’ clinical re evance rema ns unc ear (RR 10 5" 9~°/~ ~ 1 a4 7~ am,\ azole ( , ...... - ..... ~ observational cohort ’ I study 20. McCarl Hip fracture 1-5 yr of PPI use may be assoc ated with increased r sk of h p fracture but clinical proton Evidence level B, significance is still unclear case-control studies I 21. Early c 1 yr of PPI use (adjusted OR, 1.44; 95% CI, 1.26-1.544~) (OR, 1.60; 95% CI, 1.25-2.04’~) ed as I Safety/ . z 5 yr of PPt (adjusted OR 1.62; 95% CI 1.02-2 58"r) mation Vitamin Bt2 Current, long-term (> 12 me) PPI use may be associated with B!2 deficiency; Ev dence level B, 22. Small [ deficiency ~) ° clinical significance remains uncerta n (OR, 4.45 95% CI 1 47-13.34 case-control studies ! (OR, 1.82; 95% CI, 1.08-3.09~) ~ " JClinl 23. Siller-I~ PPI = proton pump inhibitor: CDAD = C. difficile-associated diarrhea: CAP = community-acquired pneumo,,~ia: OR = odds ratio: CI = confidence zole o interval: H2RA = histamine2-receptor antagonist; IgE = immunoglobulin E; RR = relative risk. Publisf 24. Gilard action Overall, data suggest that while it would be extremely (Ome~ bly IgE-mediated allergic reactions (Table III). Such associa- 25. Gilard costly and possibly unnecessary to monitor B12 levels in all tions appear modest, yet data are limited by study size and/or action patients taking acid-suppression therapy, it may be prudent design. Further studies are needed to explore the itrue rela- Publisl to monitor levels in patients who are chronically taking acid- 26. He MF tionship of PPIs to these outcomes. I~ comita suppressive therapy or in those already at higher risk for BI2 syndr( deficiency, such as the elderly or those with dementia2 27. Ray W The authors report no relevant financial relationships. and pr 28. Stockl Conclusion References clopidl 29. Aseed Understanding the possible harms of commonly pre- 1. Top 200 generic drugs by retail dollars in 2007, http://drugtopics.modern pump scribed medications such as PPIs becomes paramount, par- medicine.com/drugtopics/data/articlestandard//drugtopics/102008/500218/ talized ticularly in the elderly, where issues such as polypharmacy article.pdf. Accessed June 21, 2010. 12, 20 2. Mat Saad AZ, Collins N, Lobe MM, O’Connor HJ. Proton pump inhibitors: A survey 30. Yearsl and medication adverse effects and interactions are com- of prescribing in an Irish general hospital. Int J Clin Pract 2005;59:31-34. factor monly encountered. PPI use is clearly indicated for certain 3. Naunton M, Peterson GM, Bleasei MD. Overuse of proton pump inhibitors. J Clin 2006;; Pharm Ther2000;25:333-340. conditions such as active peptic ulcer disease or stress ulcer 4. Nardino R J, Vender R J, Herbert PN. Overuse of acid-suppressive therapy in hospi- prophylaxis in high-risk patients. However, when these con- talized patients. Am J Gastroentero12000;95 :3118-3122. 5. Parente F, Cucino C, Ga!lus S, eta!. Hospital use of acid-suppressive medications ditions are not present, discontinuation of PPIs should be and its fall-out on prescribing in general practice: A 1-month survey. Ailment considered.4’~ While it is clear that PPI therapy is beneficial Pharamacol Ther 2003; ! 7:1503-1506. when used appropriately in the treatment of various GI dis- 6. Walker NM, McDonald J. An evaluation of the use of proton pump inhibitors. Ptrarm World Sci2001;23:116-117. orders, growing evidence suggests that PPI therapy may be 7. Pohland C J, Scavnicky SA, Lasky SS, Good CB. Lansoprazole overutilization: associated with several adverse outcomes, including C. Methods for step-down therapy. Am J Manag Care 2003;9:353-358. 8. Shams D, Siddiqui NH, Heif, MM. Gastroesophageal reflux disease in older adults: cile infections, CAP, hip fractures, B12 deficiency, and possi- What is the difference? Clinical Geriatrics 2009; 17(3):32-37.

30 Clinical Geriatrics July/August 2010 9. Valuck R J, Ruscin JM. A case-control study on adverse effects: H2 blocker or pro- 31. Dial S, Alrasadi K, Manoukian C, et al. Risk of clostridium difficile diarrhea among ton pump inhibitor use and risk of vitamin B12 deficiency in older adults. J Clin hospital inpatients prescribed proton pump inhibitors: Cohort and case-control Epidemio12004 ;57: 422-428. studies. CMAJ 2004;171:33-38. 10. Vanderhoff BT, Tahboub RM. Proton pump inhibitors: An update. Am Fam Physician 32. Leonard J, Marshall JK, Moayyedi F~ Systematic review of the risk of enteric infec- 2002;66:273-280. tion in patients taking acid suppression. Arn J Gastroentero12007;102:2047-2056. 11. Pham CQ, Regal RE, Bostwick TR, et aL Acid suppressive therapy use on an inpa- Published Online: May 17, 2007. tient internal medicine service. Ann Pharmacother 2006;40:1261-1266. 33. Shah S, Lewis A, Leopold D, et al. Gastric acid suppression does not promote 12. Shi S, KIotz U. Proton pump inhibitors: An update of their clinical use and pharma- clostridial diarrhoea in the elderly. QJM2000;93:175-181. cokinetics. Eur J Clin Pharmacol 2008;64:935-95!. Published Online: August 5, 34. Cunningham R, Dale B, Undy B, Gaunt N. Proton pump inhibitors as a risk factor 2008. for clostridium difficile diarrhoea. J Hosp Infect 2003;54:243-245. 13. Detsky ME, Juudink DN Does gastric acid suppression increase the risk of com- 35. Lowe DO, Mamdani MM, Kopp A, et aL Proton pump inhibitors and hospitalization munity-acquired pneumonia? CMAJ 2005;172(3):331. for ctostridium difficile-associated disease: A population-based study. Clin Infect Dis 14. Bhatt DL, Scheiman J, Abraham NS, et al; American College of Cardiology 2006;43:1272-1276. Published Online: October 13, 2006. Foundation Task Force on Clinical Expert Consensus Documents. ACCF/ACG/AHA 36. Howell MD, Novack V, Grgurich P, et aL latrogenic gastric acid suppression and the 2008 expert consensus document on reducing the gastrointestinal risks of risk of nosocomia] Clostridiurn difficile infection. Arch Intern Med 20 ! 0; 170(9):784- antiplatelet therapy and NSAID use: A Report of the American College of 790. Cardiology Foundation Task Force on Clinical Expert Consensus Documents. 37. Untersmayr E, Bakos N, Scholl ~, et al. Anti-ulcer drugs promote IgE formation Circulation 2008; 118:1894-1909. Published Online: October 3, 2008. toward dietary antigens in adult patients. FASEB 2005;19(6):656-658. Published 15. Manoha.ran T, Darzins Ft Acid suppression in the elderly. Australasian Journal on online January 25, 2005. Ageing 2006;25(3): 164-166. 38. Untersmayr E, Sch611 I, Swoboda I, et at. Antacid medication inhibits digestion of 16. Hajjar ER, Cafiero AC, Hanlon JT. Polypharmacy in elderly patients. Am J Geriatr dietary proteins and causes food allergy: A fish allergy model in BALB/c mice. J Pharmacother2007;5(4):345-351. Allergy Clin Immuno12003; 112:616-623. 17. Sutfin T, Balmer ,K, Bostr6m H, et aL Stereoselective interaction of omeprazole with 39. Laheij R J, Sturkenoboom MC, Hassing R J, et al. Risk of community-acquired pneumonia and use of gastric acid-suppressive drugs. JAMA 2004;292(16):!955- wafarin iri healthy men. TherDrug Monit1989;11:176-184. 18. Kiley cA, Cragin D J, Roth BJ. Omeprazole-associated digoxin toxicity. South Med 1960. J 2007; 100(4):400-402. 40. Gulmez SE, Holm A, Frederiksen H, et aL Use of proton pump inhibitors and the 19. Oosterhuis 8, Jonkman JH, Anderson T, et aL Minor effect of multiple dose omper- risk of community-acquired pneumonia: A population-based case-control study. azole on the pharmacokinetics of digoxin after a single oral dose. B J Clin Arch Intern Med 2007; 167:950-995. Pharmaco11991 ;32:569-572. 41. Sarkar M, Hennessy S, Yang Y. Proton-pump inhibitor and the risk for community- 20. McCarthy DM, McLaughlin TP, Griffis DL, Yazdani C. Impact of cotherapy with some acquired pneumonia. Ann Intern Med 2008;149:391-398. proton pump inhibitors on medical claims among HMO patients already using otl~er 42. Herzig S J, Howell MD, Ngo LH, Marcantonio ER. Acid-suppressive medication use common drugs also cleared by cythochrome. Am J Ther2003;10:330-340. and the risk for hospital-acquired pneumonia. JAMA 2009;301 (20):2120-2!28. 21. Early commun~ation about ongoing safety review of clopidogrel bisulfate (market- 43. Sataloff R. Community-acquired pneumonia and use of gastric acid-suppressive ed as Plavix). U.S. Food and Drug Administration. http://www.fda.gov/Drugs/Drug drugs. JAMA 2005;293:7:795-796. SafetyiPostmarket DrugSafetylnfor mationforPatientsandProviders/DrugSafetylnfor 44. Gregor JC. Acid suppression and pneumonia: A clinical indication for rational pre- mationforHeathcareProfessionals/ucm079520.htm. Updated February 23, 2010. scribing. JAMA 2004;292:16:2012-2013. Accessed June 28, 2010. 45. Yang YX, Lewis, JD, Epstein, S. Metz DC. Long-term proton pump inhibitor therapy 22. Small DS, Farid NA, Payne CD, et al. Effects of the proton pump inhibitor lansopra- and risk of hip fracture. JAMA 2006;296:2947-2953. zote on the pharmacokinetics and pharmacodynamics of prasugrel and clopidogrel. 46. Vestergaard P, Rejnmark L, Mosekilde L, Proton pump inhibitors, histamine H2 J Clin Pharmaco12008;48:475-484. Published Online: February 26, 2008. receptor antagonists, and other antacid medications and the risk of fracture. Calcif 23. Siller-Matula JM, Spiel AO, Lang IM, et aL Effects of pantoprazole and esomepra- Tissue Int2006;79:76-83. Published Online: August 15, 2006. zole on platelet inhibition by clopidogrel. Am Heart J 2009;157:148e1-148.e5. 47. Targownik LE, Lix LM, Metge C J, et aL Use of proton pump inhibitors and risk of Published Online: November 6, 2008. osteoporosis-related fractures. CMAJ 2008; 179:319-326. 24. Gilard M, Arnaud B, Cornily JC, et al. Influence of omeprazole on the antiplatelet 48. Roux C, Bdot K, Gossec L, et aL Increase in vertebral fracture risk in post- action of clopidogrel associated with aspirin: The randomized, double-blind OCLA menopausal women using omeprazole. Calcif Tissue Int 2009;84(1):13-19. (Omeprazole CLopidogrel Aspirin) study. J Am Coil Cardio12008;51:256-260. Published Online: November 21, 2008. ssocia- 25. Gitard M, Arnaud B, Le Gal C, et aL influence of omeprazole on the antiplatelet 49. Cot6 GA, Howden CW. Potential adverse effects of proton pump inhibitors. Curr and/or action of clopidogrel associated to aspirin. J Thromb Haemost 2006;4:2508-2509. Gastroenterol Rep 2008; 10:208-214. Published Online: August 8, 2006. 50. Gray SL, LaCroix AZ, Larson J, et al. Proton pump inhibitor use, hip fracture, and !e rela- 26. Ho MP, Maddox TM, Wang L, et aL Risk of adverse outcomes associated with con- change in bone mineral density in postmenopausal women: Results from the comitant use of clopidogrel and proton pump inhibitors following acute coronary Women’s Health Initiative. Arch Intern Med 2010;170(9):765-771. syndrome JAMA 2009;301:937-944. 51. Johnson DA. Do ppis increase risk for hip fracture? Journal Watch. http://gastro 27. Ray WA, Murray KT, Griffin MR, et aL qutcomes with concurrent use of c!opidogrel enterology.jwatch.org/cgi/content/full/2007/105/l. Accessed June 21, 2010. and proton pump inhibitors: A cohort study. Ann Intern Med2010;152 (6):337-345. 62. Fried M, Siegdst H, Frei R, et aL Duodenal bacterial overgrowth during treatment in 28. Stockl KM, Le L, Zakharyan A, et aL Risk of rehospitalization for patients using outpatients with omeprazole. Gut 1994;35:23-26. clopidogrel with a proton pump inhibitor. Arch Intern Med 2010;170(8):704-710. 53. Howden CW. Vitamin B12 levels during prolonged treatment with proton pump 29. Aseeri M, Schroeder T, Kramer J, Zackula R. Gastric acid suppression by proton inhibitors. J Clin Gastroentero12000;30(1):29-33. 5 modern pump inhibitors as a risk factor for clostridium difficile-associated diarrhea in hospi- 54. Laine L, Ahnen D, Mcclain C, et aL Review article: Potential gastrointestinal effects of long-term acid suppression with proton pump inhibitors. Aliment Pharmacol Ther 500218/ talized patients. Am J Gastroentero12008;103:2308-2313. Published Online: August 12, 2008. 2000;14(6):651-668. ~ survey 30. Yearsley KA, Gilby L J, Ramadas AV, et aL Proton pump inhibitor therapy is a risk 55. Force RW, Meeker AD, Cady PS, et aL Ambulatory care increased vitamin B12 factor for clostridium difficile-associated diarrhea. Aliment Pharmacol Ther requirement associated with chronic acid-suppression therapy. Ann Pharmacother 2003;37:490-493. rs~ J Clin 2006;24:6! 3-619.

dications Aliment

~ Pharm

)r adults:

Clinical Geriatrics July/August 2010 31